Most “AI-powered” lead tools resell you tokens at a markup. You pay per credit, per row, per enrichment - and the platform keeps the spread. DataMint takes the opposite bet: bring your own keys (OpenAI, Serper, Reoon) and pay cost.
What enrichment actually costs
Enriching one row - finding a contact email, detecting phone type, scoring intent - typically takes a small model a few thousand tokens. At raw API pricing that is a fraction of a cent.
| Model | Typical per-row cost | 10,000 rows |
|---|---|---|
| Credit-based SaaS | $0.05 - $0.25 | $500 - $2,500 |
| DataMint + your key | ~$0.001 | ~$10 |
The markup is the product. Remove it and enrichment stops being a budget decision.
#Column prompts: enrichment you can read
DataMint’s enrichment agent works on your column references. Type a prompt like:
#Website → find the best contact email and note the source page
and the agent crawls, reasons, and writes results per row - showing its work in a reasoning trace you can audit. No black boxes.
Why we chose BYOK
- Cost control - you see every cent of spend in your own provider dashboard
- Model freedom - upgrade or swap models the day better ones ship
- Privacy - your prompts and rows go from your machine to your provider, not through our servers
Bring your keys once, and every enrichment run prices at cost - forever.